Two-Step Mechanism of Induction of the Gene ... - ACS Publications

Mar 25, 2008 - UniVersity of Dundee, Dundee DD1 9SY, U.K.. ReceiVed August 9, 2007. Cancer-preventive activity by exogenous molecules can be mediated ...
0 downloads 0 Views 557KB Size
Chem. Res. Toxicol. 2008, 21, 805–812

805

Articles Two-Step Mechanism of Induction of the Gene Expression of a Prototypic Cancer-Protective Enzyme by Diphenols René V. Bensasson,*,†,‡ Vincent Zoete,‡,§ Albena T. Dinkova-Kostova,‡,⊥,# and Paul Talalay‡,⊥ Laboratoire de Chimie des Substances Naturelles, MNHN, USM 0502/UMR 5154 CNRS, Case 54, 63 rue Buffon, F 75005 Paris, France, Swiss Institute of Bioinformatics, Molecular Modeling Group, Quartier Sorge, Batiment Genopode, CH-1015, Lausanne, Switzerland, Department of Pharmacology and Molecular Sciences, Johns Hopkins UniVersity School of Medicine, Baltimore, Maryland 21205, and Biomedical Research Centre, UniVersity of Dundee, Dundee DD1 9SY, U.K. ReceiVed August 9, 2007

Cancer-preventive activity by exogenous molecules can be mediated by enhancing the expression of cytoprotective enzymes [e.g, glutathione-S-transferase (GST) or NAD(P)H-quinone oxidoreductase 1 (NQO1)] via antioxidant-response elements (AREs) present in the promoter regions of their genes. Previously, potency of induction of NQO1 has been linearly correlated with the ability to release an electron from different classes of inducers, including diphenols, phenylpropenoids, and flavonoids. In the present work, we focus on the induction of NQO1 by diphenols, which we consider as a model underlying the mechanisms of action of other phenolic inducers such as phenylpropenoids and flavonoids. A two-step mechanism of NQO1 activation is proposed involving (i) oxidation of diphenol inducers to their quinone derivatives and (ii) oxidation of two highly reactive thiol groups by these quinones of a protein involved in NQO1 induction. These two putative routes are supported by linear correlations between the inducer potencies and the redox properties of diphenols and of their corresponding quinones. The linear correlations demonstrate the possibility to predict the enhanced gene expression of enzymatic defenses by diphenols from quantum mechanical calculations (i) of the ability of diphenols to release electrons and (ii) of the electron affinity of their corresponding quinones. Introduction Oxidative stress, a constant hazard to aerobic life, is generated by reactive oxygen species (ROS) arising from respiration, immune responses to infection, interaction with xenobiotics such as drugs, tobacco, alcohol, asbestos, and metals, or irradiation by visible, UV, or ionizing radiations (1, 2). Accumulation of these oxidative stresses leads to oxidation of DNA, proteins, and lipids, tissue degeneration, mutagenicity, aging, and carcinogenesis (2). Living systems have developed multiple lines of defense against these oxidative stresses. Prominent among these protective mechanisms is a family of phase 2 enzymes that are highly inducible and protect cells not only against the widespread damaging effects of ROS but also against toxicities of electrophiles and the damaging effects of inflammation. NADP(H)quinone oxidoreductase 1 (NQO1) is a prototype phase 2 enzyme that is induced coordinately with other phase 2 proteins and has played a very useful role in the assessment of the potencies of phase 2 inducers, in the discovery and isolation of * Corresponding author. Tel: +33 (01) 40 79 36 92. Fax: +33 (01) 40 79 31 35 and +33 (01) 40 79 31 47. E-mail: [email protected]. † MNHN. ‡ All authors contributed equally to this work. § Swiss Institute of Bioinformatics. ⊥ Johns Hopkins University School of Medicine. # University of Dundee.

new inducers from natural sources, and in elucidating the chemistry of inducers (3–5). Induction of NQO1 is regulated by the Keap1-Nrf2-ARE system. Certain inducers oxidize two highly reactive cysteine residues of the sensor protein Keap1, resulting in disulfide formation and conformation change, which allows Nrf2 to undergo nuclear translocation and binding to the antioxidant response element (ARE). The ARE, also known as the electrophile response element (EpRE), is represented by conserved upstream regulatory sequences that are present on many genes encoding proteins with various cytoprotective functions. After recruitment of the basic transcriptional machinery, the ultimate result is activation of the transcription of NQO1 and other AREregulated genes (6–10). One of the earliest clues to the structural requirements for phase 2 inducer activity was the finding that among diphenols (DP; see structures in Figures 2 and 3), only 1,2 diphenols (catechols) and 1,4-diphenols (hydroquinones) but not 1,3diphenols (resorcinols) were inducers (11). This finding established that oxidative lability among these compounds was essential for inducer activity but did not distinguish whether the redox process or the quinone products were the actual inducers. This issue was resolved by the subsequent realization that many inducers, such as quinones, were electrophilic Michael reaction acceptors (12). Collectively, these observations pointed to the universal reactivity of all ultimate inducers through their

10.1021/tx7002883 CCC: $40.75  2008 American Chemical Society Published on Web 03/25/2008

806

Chem. Res. Toxicol., Vol. 21, No. 4, 2008

Bensasson et al.

Figure 1. Log of the rate constant for the one-electron oxidation of reactant R by CF3O2• peroxyl radical plotted vs the one-electron oxidation potential of R (data from Table 2). Figure 4. Relative induction of NQO1 activity (by 10 µM diphenols) measured in murine Hepa 1c1c7 cells plotted as a function of the electron affinity of the quinone Q expressed by the E(HOMO) of the corresponding anion Q•- (data from Table 4).

Figure 2. Log of the rate constant k∆ for reaction of singlet oxygen 1 O2 with DP plotted vs the E(HOMO) of DP (data from Table 3).

phenolic inducers such as PP and F. A two-step mechanism of activation of Nrf2 is examined; it involves (i) formation of oxidized forms of the DP inducers and (ii) oxidation of two neighboring thiol groups of Keap1 by these oxidized species, which we propose are quinone derivatives Q (see structures in Figure 4). These two putative routes are supported by quantitative structure–activity relationships correlating NQO1 induction by DP with redox properties (i) of DP and (ii) of DP oxidized species. The redox properties examined are (i) for the first step the one-electron oxidation potentials of the DP, E(DP•+/DP) and (ii) for the second step the one-electron reduction potentials, E(Q/Q•-), of the quinone derivatives Q. Both redox properties were quantified by molecular orbital calculations.

Materials and Methods

Figure 3. Relative induction of NQO1 activity (by 10 µM diphenols) measured in murine Hepa 1c1c7 cells plotted as a function of the E(HOMO) of the DP (data from Table 3).

reduction by thiol groups and led to the recognition that cysteine residues of Keap1 are conclusive cellular sensors that react with inducers, thereby signaling phase 2 induction (13). The potency for induction of NQO1 has been linearly correlated with the ability to release an electron among different classses of inducers including diphenols (DP) (14), phenylpropenoids (PP) (15), and flavonoids (F) (16). All NQO1 inducers, DP, PP and F, in their reduced form, are unable to oxidize thiols. Thus, the only possible route for this activation is a process involving an oxidized form of these inducers. The aim of the present work was to find the role played by different species involved in the redox cascade of reactions leading to Nrf2 activation. Induction of NQO1 by diphenols (DP), measured in Hepa 1c1c7 cells, is discussed, with the possibility that the route leading to Nrf2 activation by DP is a model for the mechanism underlying activation by other

Determination of the Redox Properties of the Different Species, S, under Study. Species S are neutral molecules or anions for which we need to know the one-electron reduction potentials of the couples (S•+/S) (oxidized form/reduced form). The one-electron reduction potential, E(S•+/S), characterizes the ease of oxidation of S to S•+, as well as the ease of reduction of S•+ to S and thus represents also the electron affinity of S•+. Species S under study, that is, reduced forms of (S•+/S) couples, are the following molecules: H2O, ROH, ROOH, H2O2, vitamin C, hydroquinone, p-metoxyphenol, phenol, resorcinol, catechol, t-butyl-hydroquinone, and 3,5-t-butylcatechol. Species S are also anions, including the superoxide anion O2•-, the reduced form of singlet oxygen 1O2, and Q•-, the reduced form of quinones Q, such as 1,2-benzoquinone, 1,4-benzoquinone, t-butyl-1,4-benzoquinone, and 3,5-di-t-butyl-1,2-benzoquinone. For most of these molecules, reliable one-electron reduction potentials, E7(S•+/S), at pH 7 are already known from pulse radiolysis determinations (18). However, in the absence of reliable one-electron reduction potentials for some DP such as resorcinol, catechol, t-butylhydroquinone, or 3,5-t-butylcatechol, we have used physicochemical parameters that are linearly correlated with E7(DP•+/ DP). These redox properties are (i) a kinetic parameter, namely, the log of the rate constant (k∆) of the reaction of singlet oxygen (1O2) with DP, which we had previously determined (14), and (ii) the energy of the highest occupied molecular orbital E(HOMO), currently calculated by a simple, semiempirical quantum mechanical method, which represents the energy required to detach an electron from a molecule in the dilute

Redox Ranking of NQO1 Inducers

gas phase. In a series of related compounds S, the log of the rate constant k∆(1O2+S) (14) and the E(HOMO) of S (17) can both be linearly correlated with the one-electron oxidation potential of S, E(S•+/S), in solution. The electron affinity of quinones Q (1,2-benzoquinone, 1,4benzoquinone, t-butyl-1,4-benzoquinone, or 3,5-di-t-butyl-1,2benzoquinone), which represents the ease of reducing Q or oxidizing Q•-, is the energy liberated when an electron adds to the molecule Q in the gas phase. It is linearly related to the reduction potential, E(Q/Q•-), measured in solution and correlated also with the energy of the lowest unoccupied molecular orbital E(LUMO) of Q (17). As will be developed later, the calculation of the E(HOMO) of the Q•- anions can be considered as a more precise way to calculate the electron affinity of the neutral Q than the calculation of the energy of the lowest unoccupied molecular orbital, E(LUMO), of Q. It is important to emphasize that although the redox potentials can be strongly affected by the surrounding medium, linear correlations are observed between E(HOMO) of a series of related species S calculated in the gas phase and their ionization potential measured in the gas phase or their one-electron oxidation potential E(S•+/S) measured in different solvents. These correlations were previously observed and discussed for series of substituted derivatives of aromatic hydrocarbons and of phenolic compounds (17). Similarly, in biological quantitative structure–activity relationships (QSAR) where physicochemical properties of a series of compounds are determined ex ViVo and their biological activities measured in ViVo or in cellular systems, it is common to find succcessful linear correlations when these studies are carried out on exogenous congeners of the same general structure (18). In current IUPAC conventions, the oxidation potential of S, E(S•+/S), should now be called the reduction potential of S•+. However, the old expression “oxidation potential” of S is still often used by literature data reported here. For calculations of the E(HOMO) of S or the E(HOMO) of Q•-, we used the semiempirical AM1 quantum mechanical calculations, carried out with the Hyperchem 7.51 program. The restricted Hartree–Fock (RHF) formalism was used to complete these calculations. The conformation of the molecules was minimized by using the Polack-Ribiere minimization algorithm, until the root-mean-square of the energy gradient reached a value of 0.01 kcal/(mol Å) (15). Induction of NAD(P)H-Quinone Reductase (NQO1). Briefly, NQO1 induction was determined in Hepa 1c1c7 cell lines by measuring the rate of reduction of 2,6-dichloroindophenol by NADH reduced per minute per milligram of protein, according to the method of Prochaska et al. (11). Inductions of quinone reductase by DP were expressed as ratios of treated to controls. The choice of the particular cell line was determined by (i) its robust responsiveness to inducers, which allows quantitative comparisons of inducer potencies, and (ii) the fact that response to inducers in this cell line mimics closely the response to inducers that is observed in ViVo in a number of murine tissues (3). Statistical Calculations and Graph Plotting. Statistical calculations and graph plotting were performed using the Kaleidagraph software (version 3.6). This software calculates the linear curve fits, using the least-squared error method. The probability p-values were calculated using the standard statistical functions of Microsoft Excel 2003.

Chem. Res. Toxicol., Vol. 21, No. 4, 2008 807

Results and Discussion In a previous article (14), we had observed that the potency of NQO1 induction is higher when the electron-donating efficacy of DP inducers are larger, thus when the inducer DP is a better reducing species. The electron-donating efficacy was assessed Via the log of the rate constant (k∆) of reaction of singlet oxygen (1O2) with DP (14). However, enzyme NQO1 induction is triggered by the inducer acting in its reaction with two thiol groups of the protein Keap1 not as a reducing species but as an oxidant (13). To become an oxidant, the DP inducer must become oxidized. For this reason, the focus of the present investigation is to establish the possibility of a two-step mechanism of NQO1 induction involving (i) the formation of oxidized forms of DP inducers, ultimately their corresponding quinones Q and (ii) the oxidation of two highly reactive thiol groups in a protein by Q, which activates the NQO1 induction. In part A, we examined the biochemical reactions that can drive the first step, DP oxidation and Q formation. Our aim is to find DP oxidation reactions that can depend on the redox potentials of the phenolic derivatives. Afterward, we will examine the reactions involved in a second step, where the Q oxidize thiol groups of the sensor protein Keap1. In part B, we will try to establish correlations between the enhanced gene expression of NQO1 determined in cellular experiments and two parameters quantified by molecular orbital calculations: (i) the ability of DP to release electrons determined by their E(HOMO) with the enhanced NQO1 activity and (ii) the electron affinity of their corresponding Q determined by the E(HOMO) of their anions Q•-. A. Reactions Involved in the Activation of NQO1 by DP. We examine the two-step mechanism leading to NQO1 activation: (i) the reactions leading to oxidations of the DP inducers in cellular systems and (ii) the reactions of the DP oxidized forms with the thiol groups of the sensor protein Keap1, ultimately leading to nuclear translocation of the transcription factor Nrf2. 1. First Step: Biochemical or Enzymatic Oxidations of the DP Leading to Phenoxyl Radicals XΦO• and Quinones Q. As shown in the reactions of the catechol chosen as an example, two consecutive oxidations of DP (which are reduced quinones QH2) can lead to the formation of phenoxyl radical XΦO• and to quinone derivatives Q via equilibria 1. Let us consider some of the biochemical reactions that could lead to the oxidation of DP.

(1)

a. Oxidation of DP by Reactive Oxygen Species (ROS). One possible route of DP oxidation might occur via an attack by the radicals HO•, RO•, ROO•, and singlet molecular oxygen, 1 O2(1∆g), species involved in oxidative stress, produced physiologically by aerobic metabolism, and often called reactive oxygen species (ROS). Their redox potentials at pH 7 are reported in Table 1. Comparing the one-electron reduction potentials at pH 7, E7, of the ROS involved in oxidative stress (Table 1) with those of the DP (Table 3), we note that the E7 values of the active DP are below those of the ROS. Thus, using these E7 values as a rough guide, we can predict that equilibrium 2, shown below,

808

Chem. Res. Toxicol., Vol. 21, No. 4, 2008

Bensasson et al.

Table 1. Reduction Potentials of ROS Involved in Oxidative Stress E7 (mV)a

couples involved in oxidative stress •

+

2730b 1600c 1060d 770-1400c 650b

HO , H /H2O RO•, H+/ROH (aliphatic alkoxyl radicals) HOO•, H+/H2O2 ROO•, H+/ROOH (alkyl peroxyl radicals) 1 O2(1∆g)/O2•-

a The E7 “reduction potentials” (at pH 7) reported for different couples involved in oxidative stress are equal to the “oxidation potentials” of the respective reduced species, H2O, ROH, H2O2, ROOH, and O2•- of the different couples. b Data from ref 19. c Data from ref 20. d Data from ref 21.

Table 2. Rate Constants for Reactions of Halogenated Peroxyl Radicals reactant R ascorbic acid hydroquinone p-methoxyphenol phenol

E7 log k log k log k (mV)a (CF3O2• + R)b (CCl3O2• + R)b (CBr3O2• + R)b 300 460 600 900

8.83 7.90 7.72 6.30

8.20 7.0 6.53